Hackers don’t just go after email and social media accounts anymore. Popular AI platforms such as ChatGPT, Claude, and Perplexity can also be targeted, and if someone gets into your account, they may be able to view past chats or use your access without permission. TechCrunch’s latest guide explains how to check for suspicious activity on each service and what to do if you think your AI platform account has been compromised.
Why AI platform account security matters
AI accounts can contain sensitive conversations, personal prompts, work notes, and other data that users may not want exposed. Like any online service, these accounts are only as safe as the login protections around them. TechCrunch recommends the basics: use a unique password stored in a password manager where passwords exist, and turn on multi-factor authentication, or MFA, whenever the platform offers it.
That second step is especially important because a stolen password alone may not be enough for an attacker to log in. In the case of the three platforms covered here, ChatGPT and Perplexity offer MFA, while Claude does not use passwords at all and instead sends a login link to your email address.
How to check whether your ChatGPT account has been hacked
Open ChatGPT in a desktop browser and click your username in the bottom-left corner. From there, go to Settings, then Security and Login, and open Active Sessions. This shows where your account is currently signed in.
If you spot a device you do not recognize, you can log out of that individual session or choose Log out all to end every active login. TechCrunch notes that if you want to change your password after that, you first need to sign out of your account.
To reset the password, return to ChatGPT’s login page, click Log in in the bottom-left corner, enter your email address, select Forgot password, and continue. ChatGPT will send a six-digit code by email. Enter that code on the login page, then create a new password. You can also use the reset your password link in the email for the official instructions.
How Claude handles suspicious logins
Claude’s process is similar in spirit but different in execution because Anthropic does not use passwords for the service. To review active access, open Claude in a browser, click your username in the bottom-left corner, select Settings, and then open Account. There you’ll find the list of Active sessions.
If you see a session you do not recognize, hover over it and use the three vertical dots on the right to choose Log out or Terminate. You can also select Log out of all devices if you want to end every session at once.
After that, you log back in with your email address. Claude sends a login link to that email, so there is no password to change. That makes access control simpler in one sense, but it also means protecting your email account is especially important.
What to do if your Perplexity account looks suspicious
Perplexity works a little differently from both ChatGPT and Claude. TechCrunch says the AI-powered search engine does not show users where they are logged in, so there is no active-session list to inspect.
If you’re concerned someone may have broken into your account, open Perplexity in a browser, click your username in the bottom-left corner, and choose All settings. From there, click Sign out of all sessions and confirm the action. That forces every login to end.
Once you log out everywhere, you can sign back in by entering your email address. Perplexity will send a unique six-digit code by email, which you can enter on the site to log in. The company also lets you use the Sign in button in the email to complete access directly.
Signs to watch for across AI platforms
The most obvious warning sign is an active session or device you do not recognize. But even without a visible stranger logged in, unusual account behavior can still be a clue that something is wrong. If chats are missing, settings have changed, or you receive login emails you did not request, it is worth taking action quickly.
- Unfamiliar devices or active sessions
- Unexpected password reset or login emails
- Missing chats or changed account settings
- Alerts about logins you did not initiate
Best practices that help before anything goes wrong
The guide’s advice is straightforward: secure the account before you need to recover it. A password manager helps keep every password unique, and MFA adds another barrier if a password is exposed. Even on services like Claude and Perplexity, where authentication is email-based rather than password-based, protecting the email inbox tied to the account is still critical.
For users who rely on AI tools for work, research, or personal projects, it is worth checking these session settings periodically, not just after a suspected breach. A quick review can make it easier to catch unauthorized access before it turns into a larger privacy problem.
Source: Original report
Was this helpful?
Explore more: Application Audit & Review More Cybersecurity Tech News
Last Modified: August 16, 2026 at 1:52 am
0 views
